Skip to content
This repository was archived by the owner on Sep 9, 2025. It is now read-only.

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 

README.md

Algonauts 2023 Huggingface dataset

Here we include some scripts for organizing the Algonauts 2023 dataset in the Huggingface format.

1. Download the official dataset

Run download_data.sh to download the official challenge data. Make sure you have gdown installed.

bash download_data.sh

You can also copy or link the challenge data to algonauts_2023_challenge_data/ if you already have it downloaded.

2. Generate splits

Generate private splits derived from the official training split. This will generate a list of indices for each subject and split (train, val, testval) saved in npy format in derived_splits/.

python generate_splits.py

3. Generate the processed dataset

Generate a processed huggingface dataset for each split and fixed image size. Note, this may take a few hours.

splits="train val testval test"
for split in $splits; do
    python generate_dataset.py --split $split --img_size 256 --workers 16
done